This maybe a bit of an ngspice matter. What I think I want is to locate the net from knowing the hierarchy path to pin, and store data in RAW using the label related to the pin. As I understand it is the

net123

in ngspice that has the data and you can not use the device pin reference as an alias to the net. at the the moment the only solution I can see is: * intercept the export netlist function of xschem (to construct netlist as nodes, then enumerate each transistor, and find the arbitrary net name connected to each pin S G D B), save this as a mapping as alias (because maybe 2 transistors share the net, so a mapping is like

net123 m20.d m19.s <http://more.path.to|more.path.to>.device.and.pin.would.go.here

) * mutate the exported ngspice config, adding the resulting nets, such as

net123

(which in effect deduplicates, which is another banal issue ngspice seems to have) * rename/replace spice simulation config Then (for simulation): * intercept the 'run simulation' function, to run as normal * but post-process the resulting RAW file, doing the reverse, using the saved alias

m20.d aka net123

mapping * mutate the RAW file possible removing the

net123

and inserting back as

m20.d

, duplicating the data if necessary, for example

m19.s

might be connected to

net123

as well, so this will show as 2 datapoints in the RAW file, one for

m20.d

pin view and one for

m19.s

pin view Then load into xschem. Now TCL can access the same voltage data, that related to the connected

net123

but in terms of the device and pin

m20.d

Alternatively you could have xschem have a TCL procedure that can resolve

m20.d

as an alias, for

net123

and substitute 🙂 but it is unclear if possible or side effect to that idea